eess.IVNov 13, 2026

Foundation Model Fine-Tuning with Attention-based Multi-View Fusion for Multi-Task Cardiac Function Assessment

Authors: João Freitas, Warre Snaet, Jan D'hooge, Sandro Queirós

Organizations: Life and Health Sciences Research Institute, University of Minho, Portugal · Cardiovascular Imaging and Dynamics Laboratory, KU Leuven, Belgium · 2Ai - Applied Artificial Intelligence Laboratory, School of Technology, Polytechnic University of Cávado and Ave (IPCA), 4750-810 Barcelos, Portugal · Howest University of Applied Sciences, Kortrijk, Belgium

Abstract

In cardio-oncology, left ventricular ejection fraction (LVEF) is the standard biomarker for monitoring cardiac function during cancer therapy. However, its estimation from echocardiography is highly operator dependent, and clinically meaningful declines are often detected only after myocardial injury has progressed, motivating automated methods for standardized assessment and earlier detection of LV dysfunction. To address this need, we propose a multi-task framework integrating cardiac cycle-based temporal standardization, multi-view spatiotemporal feature extraction, and joint optimization for LVEF estimation and LV dysfunction classification. Apical four- and two-chamber echocardiographic sequences are temporally standardized by sampling a clip over an estimated cardiac cycle before being processed by the EchoPrime foundation model, which is fine-tuned end-to-end within the proposed architecture. View-specific features are fused through an attention-based aggregation module to generate a patient-level embedding, from which two task-specific prediction heads estimate continuous LVEF and binary LV dysfunction. To broaden the LVEF distribution encountered during training, samples from the MIMIC-IV-ECHO dataset are incorporated using a balanced sampling strategy. Evaluated on the EchoRisk-MICCAI Challenge, our method achieves an MAE of 4.003 for LVEF estimation (Task 1) and an AUC of 0.764 for LV dysfunction detection (Task 2) on the validation set, demonstrating the effectiveness of foundation model fine-tuning with attention-based multi-view fusion for automated cardiac function assessment in cardio-oncology.

Figures & tables

Explore similar work

Aug 14, 2026

A Unified DINOv2-Based Framework for LVEF Estimation, GLS Dysfunction Classification, and Early Cardiotoxicity Prediction

Left ventricular ejection fraction (LVEF) estimation (Task 1), global longitu-dinal strain (GLS)-based dysfunction classification (Task 2), and early cardi-otoxicity prediction (Task 3) provide complementary information for cardio-oncology assessment. LVEF reflects macroscopic ventricular volume chang-es as the clinical standard, whereas GLS captures subtle myocardial defor-mation, indicating subclinical cardiotoxicity before overt LVEF decline. Fur-thermore, predicting cardiotoxicity from baseline echocardiography prior to treatment enables preventive interventions at an early stage. To address these three tasks, we employ a DINOv2-based framework with task-specific adap-tation and prediction heads. Built upon a frozen foundation encoder, the framework incorporates parameter-efficient Low-Rank Adaptation (LoRA) and temporal aggregation to learn task-specialized representations, ensuring robust generalization. Crucially, during inference, it operates in a fully cycle-detection-free and phase-free manner, requiring neither cardiac cycle seg-mentation nor explicit End-Diastolic/End-Systolic (ED/ES) annotations. Ad-ditionally, we introduce an ED/ES-guided 2D/3D hybrid multi-view regres-sion model specifically to optimize Task 1. On a patient-level split containing 1,203 training videos from 237 patients and 300 validation videos from 59 independent patients, the DINOv2-based framework achieved a mean abso-lute error (MAE) of 5.03% for Task 1, an AUC-ROC of 76.48% for Task 2, and an AUC-ROC of 70.26% for Task 3. For Task 1, the specialized ED/ES-guided model further improves performance, achieving an MAE of 4.64%. This framework demonstrates the effectiveness of foundation model repre-sentations across diverse cardio-oncology tasks and the additional benefit of physiology-guided modeling for accurate LVEF estimation.
Mar 21, 2025

Echo-E3^3Net: Efficient Endocardial Spatio-Temporal Network for Ejection Fraction Estimation

Left ventricular ejection fraction (LVEF) is a primary marker of cardiac function. However, routine estimation from endocardial measurements requires manual delineation at end-diastole (ED) and end-systole (ES), a process that is time-consuming and subject to inter-observer variability. Reliable automation is especially valuable for point-of-care ultrasound (POCUS), where computational resources are limited and acquisition quality varies. We propose Echo-E3^3Net, an anatomy-guided spatio-temporal network that explicitly embeds cardiac anatomy into LVEF prediction. A dual-phase Endocardial Border Detector (E2^2CBD) uses phase-specific cross-attention to localize ED/ES endocardial landmarks and produce phase-aware landmark embeddings, while an Endocardial Feature Aggregator (E2^2FA) fuses these embeddings with global statistical descriptors of deep feature maps to refine EF regression. Training is guided by a lightweight geometric loss that uses ED and ES endocardial landmarks to regularize EF prediction. On EchoNet-Dynamic and a PSAX subset of EchoNet-Pediatric, Echo-E3^3Net attains competitive performance using only 1.55M parameters and 8.05 GFLOPs, an order-of-magnitude compute reduction versus recent baselines, supporting real-time deployment. Our code is publicly available at https://github.com/moeinheidari7829/Echo-E3Net.
Jul 1, 2026

EchoRisk: A Multicentre Echocardiography Dataset and Benchmark for Cardio-Oncology

Therapy-induced cardiotoxicity is the leading non-oncological cause of treatment interruption in breast cancer patients, yet early, automated risk stratification from routine cardiac imaging remains an unsolved problem. We present EchoRisk, the first curated, multicentre, longitudinal echocardiography dataset with explicit cardiotoxicity labels, released as the primary technical reference for the EchoRisk-MICCAI 2026 challenge. The dataset comprises 422 patients enrolled in the EU-funded CARDIOCARE prospective study across five European sites, yielding 2,159 echocardiography videos across 1,123 clinical exams acquired at up to five longitudinal timepoints, alongside a dedicated cohort of 280 patients with baseline imaging for early cardiotoxicity prediction. Three clinically grounded tasks are defined: automated estimation of left ventricular ejection fraction from cine video (Task 1), classification of LV dysfunction from longitudinal imaging (Task 2), and early prediction of therapy-induced cardiotoxicity from pre-therapy baseline echocardiography alone (Task 3). For each task we specify the evaluation protocol, primary and secondary metrics, and ranking procedure. We establish baseline performance using an R(2+1)D video backbone with LSTM aggregation trained from Kinetics-400 pretrained weights, demonstrating strong discriminative performance for cardiac functional assessment and LV dysfunction classification, while early cardiotoxicity prediction from a single pre-therapy video remains a significant open problem for the community. The dataset, evaluation code, and baseline implementations are publicly available to serve as a benchmark for further collaboration, comparison, and the creation of task-specific architectures in cardio-oncology.